PanoPlane: Plane-Aware Panoramic Completion for Sparse-View Indoor 3D Gaussian Splatting

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Qureshi, Adil, Jung, Dongki, Choi, Jaehoon, Manocha, Dinesh
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918500914692096
author Qureshi, Adil
Jung, Dongki
Choi, Jaehoon
Manocha, Dinesh
author_facet Qureshi, Adil
Jung, Dongki
Choi, Jaehoon
Manocha, Dinesh
contents We present PanoPlane, an approach for high-fidelity sparse-view indoor novel view synthesis that reconstructs closed room geometry via panoramic scene completion. Unlike perspective-based methods that generate training views from limited fields of view, PanoPlane leverages $360^{\circ}$ panoramic completion to condition the generative process on the full spatial layout. We propose Layout Anchored Attention Steering, a training-free mechanism that steers attention within the diffusion model's internal representation toward scene's detected planar surfaces at inference time. By directing each unobserved region's attention toward geometrically consistent observed content, our method replaces unconstrained hallucination with grounded surface extrapolation. The resulting panoramic completions provide supervision for 3D Gaussian Splatting, enabling accurate novel-view synthesis across unobserved regions from as few as three input views. Experiments on Replica, ScanNet++, and Matterport3D demonstrate state-of-the-art novel view synthesis quality across 3, 6, and 9 input views, achieving up to $+17.8\%$ improvement in PSNR over the current state-of-the-art baseline without any training or fine-tuning of the diffusion model.
format Preprint
id arxiv_https___arxiv_org_abs_2605_14135
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PanoPlane: Plane-Aware Panoramic Completion for Sparse-View Indoor 3D Gaussian Splatting
Qureshi, Adil
Jung, Dongki
Choi, Jaehoon
Manocha, Dinesh
Computer Vision and Pattern Recognition
We present PanoPlane, an approach for high-fidelity sparse-view indoor novel view synthesis that reconstructs closed room geometry via panoramic scene completion. Unlike perspective-based methods that generate training views from limited fields of view, PanoPlane leverages $360^{\circ}$ panoramic completion to condition the generative process on the full spatial layout. We propose Layout Anchored Attention Steering, a training-free mechanism that steers attention within the diffusion model's internal representation toward scene's detected planar surfaces at inference time. By directing each unobserved region's attention toward geometrically consistent observed content, our method replaces unconstrained hallucination with grounded surface extrapolation. The resulting panoramic completions provide supervision for 3D Gaussian Splatting, enabling accurate novel-view synthesis across unobserved regions from as few as three input views. Experiments on Replica, ScanNet++, and Matterport3D demonstrate state-of-the-art novel view synthesis quality across 3, 6, and 9 input views, achieving up to $+17.8\%$ improvement in PSNR over the current state-of-the-art baseline without any training or fine-tuning of the diffusion model.
title PanoPlane: Plane-Aware Panoramic Completion for Sparse-View Indoor 3D Gaussian Splatting
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.14135